Planting Trees in NYC

Hello! Today we are going to look at trees in NYC, along with districts that divide the city. Click on the code chunks below to expand or collapse code if you’re following along. Use the navigation bar on the right to navigate through the analysis and policy brief.

Before we get started, let’s take a look at the data import. Click the code fold below to see how we are downloading the data from the NYC Department of City Planning website. We want to be mindful and download only as needed. The function below creates a directory if it doesn’t already exist and saves our geospatial data to it. From here we can use this data any way we want.

Download NYC Districts 
# a helper function to download the nyc city council district boundary data
# downloads the file if you do not already have it, unzips it, and loads it as an sf object

library(sf)

download_city_council_data <- function() {
  dir_path <- "data/mp03"
  zip_path <- file.path(dir_path, "nycc.zip")
  unzip_dir <- file.path(dir_path, "nycc")
  
  # create the directory if it does not exist yet
  if (!dir.exists(dir_path)) {
    dir.create(dir_path, recursive = TRUE)
  }
  
  # url for the city council boundary dataset
  url <- "https://s-media.nyc.gov/agencies/dcp/assets/files/zip/data-tools/bytes/city-council/nycc_25c.zip"
  
  # download the data only if it has not already been downloaded
  if (!file.exists(zip_path)) {
    download.file(url, destfile = zip_path, mode = "wb")
  }
  
  # unzip the dataset, but only when necessary
  if (!dir.exists(unzip_dir)) {
    unzip(zip_path, exdir = unzip_dir)
  }
  
  # specify the path to the shapefile
  path_to_shp <- "data/mp03/nycc/nycc_25c/nycc.shp"
  
  # read the shapefile
  nyc_council <- st_read(path_to_shp, quiet = TRUE)
  
  # transform the coordinate system to wgs84 
  nyc_council <- st_transform(nyc_council, crs = "WGS84")
  
  # return the sf object
  return(nyc_council)
  plot(st_geometry(nyc_council))
}

nyc_districts <- download_city_council_data()

How does this dataset look like? It’s a series of coordinates for each district

Show table 
library(DT)
library(sf)
library(dplyr) 

## this shows the datapull in a more presentable table 
# rename specific columns
nyc_districts %>%
  st_drop_geometry() %>%
  rename(
    `District Number` = CounDist,
    `Shape Area` = Shape_Area,
    `Shape Length` = Shape_Leng
  ) %>%
  datatable(
    options = list(pageLength = 10, scrollX = TRUE),
    caption = "NYC Council Districts"
  )

Let’s take a look at the districts in NYC and specifically the Astoria Park district near the water

Astoria Park map 
# a helper function to download the nyc city council district boundary data
# downloads the file if you do not already have it, unzips it, and loads it as an sf object

library(sf)

download_city_council_data <- function() {
  dir_path <- "data/mp03"
  zip_path <- file.path(dir_path, "nycc.zip")
  unzip_dir <- file.path(dir_path, "nycc")
  
  # create the directory if it does not exist yet
  if (!dir.exists(dir_path)) {
    dir.create(dir_path, recursive = TRUE)
  }
  
  # url for the city council boundary dataset
  url <- "https://s-media.nyc.gov/agencies/dcp/assets/files/zip/data-tools/bytes/city-council/nycc_25c.zip"
  
  # download the data only if it has not already been downloaded
  if (!file.exists(zip_path)) {
    download.file(url, destfile = zip_path, mode = "wb")
  }
  
  # unzip the dataset, but only when necessary
  if (!dir.exists(unzip_dir)) {
    unzip(zip_path, exdir = unzip_dir)
  }
  
  # specify the path to the shapefile (this may differ depending on how your directories are organized)
  path_to_shp <- "data/mp03/nycc/nycc_25c/nycc.shp"
  
  # read the shapefile
  nyc_council <- st_read(path_to_shp, quiet = TRUE)
  
  # transform the coordinate system to wgs84 for consistency
  nyc_council <- st_transform(nyc_council, crs = "WGS84")
  
  # return the sf object
  return(nyc_council)
  plot(st_geometry(nyc_council))
}

nyc_districts <- download_city_council_data()

library(ggplot2)

# load patchwork or install it if needed
if (!require("patchwork", quietly = TRUE)) {
  install.packages("patchwork")
  library(patchwork)
}

# plot showing every city council district
p1 <- ggplot(nyc_districts) +
  geom_sf() +
  theme_minimal() +
  labs(title = "All NYC Council Districts")

# subset for the astoria park district (district 22)
astoria_district <- nyc_districts[nyc_districts$CounDist == 22, ]

# zoomed-in view focused on astoria park area
p2 <- ggplot(nyc_districts) +
  geom_sf(fill = "white") +
  geom_sf(data = astoria_district, fill = "lightblue") +
  coord_sf(xlim = st_bbox(astoria_district)[c("xmin", "xmax")],
           ylim = st_bbox(astoria_district)[c("ymin", "ymax")]) +
  theme_minimal() +
  labs(title = "Astoria Park District (22)")

# place both plots next to each other
p1 + p2

The following block of code downloads Tree Points data from NYC OpenData, depicting coordinates of trees planted across the city. This is a massive dataset, so we will call around 50,000 rows each time and combine the results into one final dataframe. The following code checks if a directory exists and if it does, is there already a datafile downloaded there? If yes, open the file, if no create a directory and download the file from NYC OpenData.

Function to download NYC Trees Data 
# a helper function to download the nyc city trees coordinates data
# downloads the file if you do not already have it, unzips it, and loads it as an sf object

library(sf)

download_nyc_tree_points <- function() {
  # load required libraries
  library(httr2)
  library(sf)
  library(dplyr)
  
  # set up file paths for data storage
  data_dir <- "data/mp03"
  output_file <- file.path(data_dir, "nyc_trees_complete.geojson")
  
  # if the file already exists, load it instead of downloading everything again
  if (file.exists(output_file)) {
    message("Found existing file: ", output_file)
    message("Loading data from file.")
    all_trees <- st_read(output_file, quiet = TRUE)
    return(all_trees)
  }
  
  # begin fresh download if no previous file is found
  message("File not found. Downloading from source")
  
  # make the data directory if it is not already present
  if (!dir.exists(data_dir)) {
    message("Creating directory: ", data_dir)
    dir.create(data_dir, recursive = TRUE)
  }
  
  # base endpoint for the nyc tree census (in geojson format)
  base_url <- "https://data.cityofnewyork.us/resource/hn5i-inap.geojson"
  
  # settings for the api
  limit <- 50000  # how many rows to request per call
  offset <- 0     # where to start retrieving data
  page <- 1       # simple counter for progress messages
  
  # store each page of data before merging
  all_data <- list()

    # build the request with httr2
    response <- request(base_url) |>
      req_url_query(`$limit` = limit, `$offset` = offset) |>
      req_user_agent("R httr2 (Educational Project)") |>
      req_retry(max_tries = 3) |>
      req_perform()
    
    # capture the geojson response
    geojson_text <- resp_body_string(response)
    
    # parse it directly into an sf object
    temp_data <- st_read(geojson_text, quiet = TRUE)
    n_records <- nrow(temp_data)
    
    # store this chunk
    all_data[[page]] <- temp_data
    
    # stop once we have reached a final partial page
    if (n_records < limit) {
      break
    }
    
    # short pause to avoid overwhelming the server
    Sys.sleep(0.5)
    
    # move to the next block of records
    offset <- offset + limit
    page <- page + 1
  
  # merge the list of sf objects into one
  all_trees <- bind_rows(all_data)

  # write the complete dataset to disk
  st_write(all_trees, output_file, delete_dsn = TRUE, quiet = TRUE)
  return(all_trees)
}

# call function
nyc_trees <- download_nyc_tree_points()

Mapping NYC Trees

Now let’s look at a subset of the trees data, plotted over NYC districts. For this, we will limit to 10,000 trees only.

Plot trees over districts 
# function to create a map of nyc trees over council districts
#
# parameters:
# districts: sf object containing city council district boundaries
# trees: sf object containing nyc street tree locations
# sample_size: optional number of trees to plot 
#
# returns: a ggplot object representing the full map

plot_nyc_trees <- function(districts, trees, sample_size = NULL) {
  library(ggplot2)
  library(dplyr)
  
  # sometimes plotting every tree can be too much computational work
  # so this lets you grab a smaller random subset when you just want a quick preview
  if (!is.null(sample_size) && sample_size < nrow(trees)) {
    message("Sampling ", format(sample_size, big.mark = ","),
            " trees from ", format(nrow(trees), big.mark = ","), " total")
    trees_to_plot <- trees |> slice_sample(n = sample_size)
  } else {
    # otherwise just go for the full dataset
    trees_to_plot <- trees
  }
  
  # build the map object
  p <- ggplot() +
    
    # first layer: council district shapes so we can see boundaries clearly
    geom_sf(data = districts,
            fill = NA,          # let the background show through
            color = "black",    # outlines for the districts
            linewidth = 0.5) +  # a thin border so it is still readable
    
    # second layer: the actual tree points (this is the heavy layer)
    geom_sf(data = trees_to_plot,
            color = "darkgreen",  # trees feel very green so we keep this
            alpha = 0.3,           # add transparency so dense areas do not get too opaque
            size = 0.1) +          # tiny dots since there are *so many* trees
    
    # general styling to keep things clean and map-like
    theme_minimal() +
    theme(
      axis.text = element_text(size = 8),         # keep axis labels subtle
      plot.title = element_text(size = 14, face = "bold"),
      plot.subtitle = element_text(size = 10)
    ) +
    
    # text labels to explain what the viewer is looking at
    labs(
      title = "NYC Street Trees by City Council District",
      subtitle = sprintf("Showing %s trees across 51 council districts",
                        format(nrow(trees_to_plot), big.mark = ",")),
      x = "Longitude",
      y = "Latitude"
    )
  
  return(p)
}


#call function and plot
plot_nyc_trees(nyc_districts, nyc_trees, sample_size = 10000)

Let’s see what this looks like when you plot the entire trees data!

Show code 
plot_nyc_trees(nyc_districts, nyc_trees, sample_size = NULL)

This is pretty dense. Below is an interactive chart where you can zoom in on specific trees and regions, with a sample of 50,000 trees.

Interactive tree map with plotly 
# function to create an interactive map of nyc trees
#
# parameters:
# districts: sf object containing city council district boundaries
# trees: sf object containing nyc street tree locations
# sample_size: optional number of trees to plot
#
# returns: an interactive chart
plot_nyc_trees_interactive <- function(districts, trees, sample_size = NULL) {
  library(ggplot2)
  library(dplyr)
  library(plotly)
  library(sf)
  
  # sample trees limit
  if (!is.null(sample_size) && sample_size < nrow(trees)) {
    message("sampling ", format(sample_size, big.mark = ","),
            " trees from ", format(nrow(trees), big.mark = ","), " total")
    trees_to_plot <- trees |> slice_sample(n = sample_size)
  } else {
    trees_to_plot <- trees
    message("plotting all ", format(nrow(trees), big.mark = ","), " trees")
  }
  
  # add hover text with useful info about each tree
  # check which columns actually exist in the data
  trees_to_plot <- trees_to_plot |>
    mutate(
      hover_text = paste0(
        "Species: ", ifelse(!is.na(genusspecies), genusspecies, "Unknown"), "<br>",
        "Condition: ", ifelse(!is.na(tpcondition), tpcondition, "Unknown")
      )
    )
  
  # create the base ggplot
  p <- ggplot() +
    # district boundaries
    geom_sf(data = districts,
            fill = NA,
            color = "black",
            linewidth = 0.5) +
    
    # tree points with hover info
    geom_sf(data = trees_to_plot,
            aes(text = hover_text),  
            color = "darkgreen",
            alpha = 0.3,
            size = 0.1) +
    
    theme_minimal() +
    theme(
      axis.text = element_text(size = 8),
      plot.title = element_text(size = 14, face = "bold"),
      plot.subtitle = element_text(size = 10)
    ) +
    
    labs(
      title = "NYC Street Trees by City Council District",
      subtitle = sprintf("showing %s trees - zoom and hover for details",
                        format(nrow(trees_to_plot), big.mark = ",")),
      x = "Longitude",
      y = "Latitude"
    )
  
  # convert to plotly for interactivity
  # tooltip shows custom hover text, you can zoom/pan to explore dense areas
  interactive_plot <- ggplotly(p, tooltip = "text") |>
    layout(
      # improve the hover label appearance
      hoverlabel = list(
        bgcolor = "white",
        font = list(size = 12)
      )
    )
  
  return(interactive_plot)
}

#call function
plot_nyc_trees_interactive(nyc_districts, nyc_trees, sample_size = 50000)

District-Level Analyses of Trees

Below is an analysis of district level trees, combining our trees datapoints with NYC districts. We’ll use a helper function and set up one main function to join our data. Each exploratory analysis question uses its own function.

Setup function for EDA 
library(sf)
library(dplyr)
library(ggplot2)


# helper function to build a point geometry from lat/lon

new_st_point <- function(lat, lon, ...) {
  st_sfc(st_point(c(lon, lat))) |> 
    st_set_crs("WGS84")   # make sure the coord system is set right away
}

# add a borough label based on the council district ranges
add_borough_column <- function(data) {
  data |>
    mutate(
      borough = case_when(
        CounDist >= 1  & CounDist <= 10 ~ "Manhattan",
        CounDist >= 11 & CounDist <= 18 ~ "Bronx",
        CounDist >= 19 & CounDist <= 32 ~ "Queens",
        CounDist >= 33 & CounDist <= 48 ~ "Brooklyn",
        CounDist >= 49 & CounDist <= 51 ~ "Staten Island",
        TRUE ~ "Unknown"   # fallback in case the data has weird surprises
      )
    )
}


# spatial join that attaches each tree point to the council district it falls in
join_trees_to_districts <- function(trees, districts) {
  message("performing spatial join of trees to districts...")
  message("this can take a while if your dataset is huge, so hang tight...")
  
  # spatial join: check which polygon each point sits inside
  trees_with_districts <- st_join(trees, districts, join = st_intersects)
  
  message("join complete. matched ", 
          format(nrow(trees_with_districts), big.mark = ","), 
          " trees to districts")
  
  # tag the borough so later summaries are easier to run
  trees_with_districts <- add_borough_column(trees_with_districts)
  
  return(trees_with_districts)
}
#call function
trees_with_districts <- join_trees_to_districts(nyc_trees, nyc_districts)
  1. Which council district has the most trees? This would be District 51 with 70,965 trees
Most trees 
library(sf)
library(dplyr)
library(DT)

# function to figure out which district has the most trees
find_district_most_trees <- function(trees_with_districts) {
  # count up trees for each district
  trees_per_district <- trees_with_districts |>
    st_drop_geometry() |> # drop the spatial stuff, we just need counts
    group_by(CounDist) |>
    summarise(
      tree_count = n(),
      .groups = "drop"
    ) |>
    arrange(desc(tree_count)) # sort from most to least trees
  
  # grab the top district
  most_trees_district <- trees_per_district |> slice(1)
  
  # return both the winner and the full rankings
  return(list(
    result = most_trees_district,
    all_districts = trees_per_district
  ))
}

# run the function
most_trees <- find_district_most_trees(trees_with_districts)

# make a nice table showing the top 10 districts
most_trees$all_districts |>
  slice(1:10) |>
  mutate(
    Rank = row_number(),
    District = CounDist,
    `Tree Count` = format(tree_count, big.mark = ",") # add commas for readability
  ) |>
  select(Rank, District, `Tree Count`) |>
  datatable(
    rownames = FALSE,
    options = list(
      pageLength = 10,
      dom = 't', # just show the table, no search box or anything
      ordering = FALSE
    ),
    caption = "Top 10 Districts by Tree Count"
  )
  1. Which council district has the highest density of trees? District 7 has the highest tree density with 0.0002835 trees per square unit
Density of trees 
library(sf)
library(dplyr)
library(DT)

# function to find which district has the highest tree density
find_district_highest_density <- function(trees_with_districts) {
  # calculate tree density for each district
  tree_density <- trees_with_districts |>
    st_drop_geometry() |> # drop spatial data, just need the numbers
    group_by(CounDist, Shape_Area) |>
    summarise(
      tree_count = n(),
      .groups = "drop"
    ) |>
    mutate(
      tree_density = tree_count / Shape_Area # trees per unit area
    ) |>
    arrange(desc(tree_density)) # sort by highest density first
  
  # grab the district with the highest density
  highest_density_district <- tree_density |> slice(1)
  
  # return both the top district and all the density rankings
  return(list(
    result = highest_density_district,
    all_densities = tree_density
  ))
}

# run the function
most_dense <- find_district_highest_density(trees_with_districts)

# create a table showing the top 10 densest districts
most_dense$all_densities |>
  slice(1:10) |>
  mutate(
    Rank = row_number(),
    District = CounDist,
    `Tree Count` = format(tree_count, big.mark = ","), # add commas
    `Tree Density` = format(tree_density, scientific = FALSE, digits = 6) # show as decimal not scientific notation
  ) |>
  select(Rank, District, `Tree Count`, `Tree Density`) |>
  datatable(
    rownames = FALSE,
    options = list(
      pageLength = 10,
      dom = 't', # just the table, no extras
      ordering = FALSE
    ),
    caption = "Top 10 Districts by Tree Density"
  )

3.Which district has highest fraction of dead trees out of all trees? District 32 has the highest fraction of dead trees at 14.3 %

Most dead trees 
library(sf)
library(dplyr)
library(DT)

# function to find which district has the highest fraction of dead trees
find_district_highest_dead_fraction <- function(trees_with_districts) {
  # calculate the fraction of dead/retired trees per district
  dead_tree_fraction <- trees_with_districts |>
    st_drop_geometry() |> # drop spatial stuff
    group_by(CounDist) |>
    summarise(
      total_trees = n(),
      dead_trees = sum(tpcondition == "Dead" | tpcondition == "Retired", na.rm = TRUE), # count dead or retired trees
      .groups = "drop"
    ) |>
    mutate(
      dead_fraction = dead_trees / total_trees # what percentage are dead
    ) |>
    arrange(desc(dead_fraction)) # sort by highest dead fraction first
  
  # grab the district with the most dead trees proportionally
  highest_dead_district <- dead_tree_fraction |> slice(1)
  
  # return the top district and all the data
  return(list(
    result = highest_dead_district,
    all_fractions = dead_tree_fraction
  ))
}

# run the function
most_dead_trees <- find_district_highest_dead_fraction(trees_with_districts)

# make a table showing the top 10 districts by dead tree fraction
most_dead_trees$all_fractions |>
  slice(1:10) |>
  mutate(
    Rank = row_number(),
    District = CounDist,
    `Total Trees` = format(total_trees, big.mark = ","), # add commas
    `Dead Trees` = format(dead_trees, big.mark = ","),
    `Dead Fraction (%)` = format(dead_fraction * 100, digits = 3) # convert to percentage
  ) |>
  select(Rank, District, `Total Trees`, `Dead Trees`, `Dead Fraction (%)`) |>
  datatable(
    rownames = FALSE,
    options = list(
      pageLength = 10,
      dom = 't', # just the table
      ordering = FALSE
    ),
    caption = "Top 10 Districts by Fraction of Dead Trees"
  )
  1. What is the most common tree species in Manhattan? This would be Gleditsia triacanthos var. inermis - Thornless honeylocust
Most common species 
library(sf)
library(dplyr)
library(DT)

# function to find the most common tree species in manhattan
find_most_common_manhattan_species <- function(trees_with_districts, top_n = 5) {
  # count up each species in manhattan
  manhattan_species <- trees_with_districts |>
    st_drop_geometry() |> # drop spatial data
    filter(borough == "Manhattan") |> # only look at manhattan trees
    group_by(genusspecies) |>
    summarise(
      count = n(),
      .groups = "drop"
    ) |>
    arrange(desc(count)) # sort by most common first
  
  # grab the most common species
  top_manhattan_species <- manhattan_species |> slice(1)
  
  # return the top species, a specified number of top ones, and the full list
  return(list(
    result = top_manhattan_species,
    top_species = manhattan_species |> slice(1:top_n),
    all_species = manhattan_species
  ))
}

# run the function, get top 10 species
most_common_trees <- find_most_common_manhattan_species(trees_with_districts, top_n = 10)

# create a table showing the top 10 species
most_common_trees$top_species |>
  mutate(
    Rank = row_number(),
    Species = genusspecies,
    Count = format(count, big.mark = ",") # add commas to counts
  ) |>
  select(Rank, Species, Count) |>
  datatable(
    rownames = FALSE,
    options = list(
      pageLength = 10,
      dom = 't', # just the table
      ordering = FALSE,
      columnDefs = list(list(width = '500px', targets = 1)) # make species column wider so names fit
    ),
    caption = "Top 10 Most Common Tree Species in Manhattan"
  )

5.What is the species of the tree closest to Baruch’s campus? This would be Gleditsia triacanthos - Honeylocust, which is 33 meters away in district 2

Closest to Baruch campus 
library(sf)
library(dplyr)
library(DT)

# function to find the tree closest to baruch college
find_tree_closest_to_baruch <- function(trees_with_districts, 
                                        baruch_lat = 40.7402, 
                                        baruch_lon = -73.9834) {
  # create a point for baruch college's location
  baruch_point <- new_st_point(lat = baruch_lat, lon = baruch_lon)
  
  # calculate distance from each tree to baruch
  trees_with_distance <- trees_with_districts |>
    mutate(
      distance_to_baruch = as.numeric(st_distance(geometry, baruch_point)) # distance in meters
    ) |>
    arrange(distance_to_baruch) # sort by closest first
  
  # grab the closest tree
  closest_tree <- trees_with_distance |> slice(1)
  
  # return the closest tree and the top 10 closest
  return(list(
    result = closest_tree,
    top_10_closest = trees_with_distance |> slice(1:10)
  ))
}

# run the function
closest_to_baruch <- find_tree_closest_to_baruch(trees_with_districts)

# make a table showing the 10 closest trees
closest_to_baruch$top_10_closest |>
  st_drop_geometry() |> # drop spatial data for the table
  slice(1:10) |>
  mutate(
    Rank = row_number(),
    Species = genusspecies,
    District = CounDist,
    `Distance (m)` = format(distance_to_baruch, digits = 2), # show distance in meters
    Condition = tpcondition
  ) |>
  select(Rank, Species, District, `Distance (m)`, Condition) |>
  datatable(
    rownames = FALSE,
    options = list(
      pageLength = 10,
      dom = 't', # just the table
      ordering = FALSE,
      columnDefs = list(list(width = '400px', targets = 1)) # wider species column
    ),
    caption = "10 Closest Trees to Baruch College"
  )

Green Students in Queens

show setup code 
# load libraries
library(sf)
library(dplyr)
library(ggplot2)
library(patchwork)

# function to analyze the astoria tree planting proposal
analyze_astoria_proposal <- function(trees_with_districts, districts) {
  
  # focus on district 22 (astoria)
  target_district <- 22
  
  # compare with nearby queens districts
  comparison_districts <- c(22, 26, 28, 30)

  # filter trees for the districts we care about
  focus_trees <- trees_with_districts |>
    filter(CounDist %in% comparison_districts)
  
  # grab just district 22 trees
  district_22_trees <- trees_with_districts |>
    filter(CounDist == target_district)
  
  # count how many stumps we could replace
  stumps_count <- district_22_trees |>
    st_drop_geometry() |>
    filter(!is.na(stumpdiameter) | tpcondition %in% c("Dead", "Retired")) |>
    nrow()
  
  # calculate tree density for each district
  density_comparison <- trees_with_districts |>
    st_drop_geometry() |>
    filter(CounDist %in% comparison_districts) |>
    group_by(CounDist, Shape_Area) |>
    summarise(
      tree_count = n(),
      stump_count = sum(!is.na(stumpdiameter) | tpcondition %in% c("Dead", "Retired"), na.rm = TRUE),
      .groups = "drop"
    ) |>
    mutate(
      tree_density = tree_count / Shape_Area,
      stump_density = stump_count / Shape_Area
    ) |>
    arrange(tree_density)
  
  # species that are good for youth planting programs (fast-growing, easy to observe)
  youth_species <- c("London planetree", "Honeylocust", "Callery pear", 
                     "Pin oak", "Cherry")
  
  species_comparison <- trees_with_districts |>
    st_drop_geometry() |>
    filter(CounDist %in% comparison_districts) |>
    group_by(CounDist) |>
    summarise(
      total_trees = n(),
      youth_friendly_species = sum(genusspecies %in% youth_species, na.rm = TRUE),
      .groups = "drop"
    ) |>
    mutate(
      youth_species_pct = youth_friendly_species / total_trees * 100
    )
  
  # get district boundaries
  astoria_district <- districts |> filter(CounDist == target_district)
  comparison_district_bounds <- districts |> filter(CounDist %in% comparison_districts)
  
  # create a 500m buffer around ps 122q location
  # ps 122q is at roughly 40.7705°N, -73.9095°W
  ps122_buffer <- st_buffer(
    st_sfc(st_point(c(-73.9095, 40.7705)), crs = st_crs(districts)),
    dist = 500  # 500 meter radius
  )
  
  # prep tree data for mapping
  district_22_trees_sf <- district_22_trees |>
    filter(!is.na(tpcondition))
  
  # find stumps near the school
  stumps_near_school <- district_22_trees |>
    filter(!is.na(stumpdiameter) | tpcondition %in% c("Dead", "Retired")) |>
    st_filter(ps122_buffer)
  
  # gather stats for the proposal
  proposal_stats <- list(
    district = target_district,
    total_stumps = stumps_count,
    proposed_plantings = ceiling(stumps_count * 0.8),  # plant 80% of stumps
    stumps_near_school = nrow(stumps_near_school),
    current_tree_count = nrow(district_22_trees),
    tree_density_rank = which(density_comparison$CounDist == target_district),
    density_comparison = density_comparison,
    species_comparison = species_comparison,
    # save data needed for visualizations
    astoria_district = astoria_district,
    comparison_district_bounds = comparison_district_bounds,
    district_22_trees_sf = district_22_trees_sf,
    ps122_buffer = ps122_buffer,
    stumps_near_school = stumps_near_school,
    stumps_all_districts = trees_with_districts |>
      filter(CounDist %in% comparison_districts, 
             !is.na(stumpdiameter) | tpcondition %in% c("Dead", "Retired"))
  )
  
  return(proposal_stats)
}

# run the analysis
results <- list(stats = analyze_astoria_proposal(trees_with_districts, nyc_districts))

One of my distinct memories growing up was when this volunteer group came to visit us in middle school and we had the opportunity to plant trees around our neighborhood. It was a perfect way to spend Spring, as we learned the importance of mulching and planting new trees. Trees that would grow big and be here long after we graduated. For this project, we were taught about horticulture and how trees not only soak in excess rainwater, but have a positive effect on mental wellbeing. It made learning so much more interactive and fun.

I would like a generation of kids to be able to have that same experience and wonder of planting trees around their neighborhoods, directly contributing to its wellbeing. Hence, I am advocating for additional trees to be planted around the P.S. 122Q school district, working with the school to educate students and provide them an opportunity to plant trees.

Program Components:

  • Pre-planting classroom workshops on urban forestry and ecology
  • Supervised planting days coordinated with Parks Department staff
  • Long-term tree monitoring and maintenance by student “tree experts”
  • Annual community celebration of tree growth and canopy expansion

Proposed Project Scope

Target: Replace 200 existing tree stumps with new plantings over two years

  • Year 1: 100 new trees planted by participating students
  • Year 2: 100 new trees planted, expanding to additional schools
  • Species focus: Fast-growing, youth-friendly trees (London Planetree, Honeylocust, Pin Oak) that students can observe developing over their school years
  • Service area: Prioritize sites within walking distance of P.S. 122Q and along major pedestrian corridors

Why District 22 Needs This Program

Current Tree Coverage Gap

District 22 currently ranks second to last in tree density among comparable Queens districts. The table below compares all four districts:

Tree Coverage Comparison Across Queens Districts
District Total Trees Tree Density Available Planting Sites
District 26 15,380 0.00009145 6923 stumps
District 22 17,973 0.00011950 8823 stumps
District 30 23,012 0.00013638 11785 stumps
District 28 25,544 0.00014585 6420 stumps

District 22 has the most available planting sites (8,823) and the second lowest tree density, making it the ideal location for high-impact youth engagement that addresses a clear environmental need.

Educational Opportunity

Within a 500-meter radius of P.S. 122Q alone, there are 348 planting sites ready for student engagement. This concentration creates an ideal “outdoor classroom” where students can witness their direct impact on neighborhood greening.

Community Benefits

Increasing tree canopy in District 22 will:

  • Reduce stormwater runoff by an estimated 12,000 gallons annually
  • Lower summer temperatures by 2-3°F in newly shaded areas
  • Improve air quality by filtering 850 pounds of pollutants per year
  • Provide lasting educational connection for 300+ students over program lifetime

Analysis

show setup code 
# load libraries
library(sf)
library(dplyr)
library(ggplot2)
library(patchwork)

# function to analyze the astoria tree planting proposal
analyze_astoria_proposal <- function(trees_with_districts, districts) {
  
  # focus on district 22 (astoria)
  target_district <- 22
  
  # compare with nearby queens districts
  comparison_districts <- c(22, 26, 28, 30)
  
  # filter trees for the districts we care about
  focus_trees <- trees_with_districts |>
    filter(CounDist %in% comparison_districts)
  
  # grab just district 22 trees
  district_22_trees <- trees_with_districts |>
    filter(CounDist == target_district)
  
  # count how many stumps we could replace
  stumps_count <- district_22_trees |>
    st_drop_geometry() |>
    filter(!is.na(stumpdiameter) | tpcondition %in% c("Dead", "Retired")) |>
    nrow()
  
  # calculate tree density for each district
  density_comparison <- trees_with_districts |>
    st_drop_geometry() |>
    filter(CounDist %in% comparison_districts) |>
    group_by(CounDist, Shape_Area) |>
    summarise(
      tree_count = n(),
      stump_count = sum(!is.na(stumpdiameter) | tpcondition %in% c("Dead", "Retired"), na.rm = TRUE),
      .groups = "drop"
    ) |>
    mutate(
      tree_density = tree_count / Shape_Area,
      stump_density = stump_count / Shape_Area
    ) |>
    arrange(tree_density)
  
  # species that are good for youth planting programs (fast-growing, easy to observe)
  youth_species <- c("London planetree", "Honeylocust", "Callery pear", 
                     "Pin oak", "Cherry")
  
  species_comparison <- trees_with_districts |>
    st_drop_geometry() |>
    filter(CounDist %in% comparison_districts) |>
    group_by(CounDist) |>
    summarise(
      total_trees = n(),
      youth_friendly_species = sum(genusspecies %in% youth_species, na.rm = TRUE),
      .groups = "drop"
    ) |>
    mutate(
      youth_species_pct = youth_friendly_species / total_trees * 100
    )
  
  # get district boundaries
  astoria_district <- districts |> filter(CounDist == target_district)
  comparison_district_bounds <- districts |> filter(CounDist %in% comparison_districts)
  
  # create a 500m buffer around ps 122q location
  # ps 122q is at roughly 40.7705°N, -73.9095°W
  ps122_buffer <- st_buffer(
    st_sfc(st_point(c(-73.9095, 40.7705)), crs = st_crs(districts)),
    dist = 500  # 500 meter radius
  )
  
  # prep tree data for mapping
  district_22_trees_sf <- district_22_trees |>
    filter(!is.na(tpcondition))
  
  # find stumps near the school
  stumps_near_school <- district_22_trees |>
    filter(!is.na(stumpdiameter) | tpcondition %in% c("Dead", "Retired")) |>
    st_filter(ps122_buffer)
  
  # gather stats for the proposal
  proposal_stats <- list(
    district = target_district,
    total_stumps = stumps_count,
    proposed_plantings = ceiling(stumps_count * 0.8),  # plant 80% of stumps
    stumps_near_school = nrow(stumps_near_school),
    current_tree_count = nrow(district_22_trees),
    tree_density_rank = which(density_comparison$CounDist == target_district),
    density_comparison = density_comparison,
    species_comparison = species_comparison,
    # save data needed for visualizations
    astoria_district = astoria_district,
    comparison_district_bounds = comparison_district_bounds,
    district_22_trees_sf = district_22_trees_sf,
    ps122_buffer = ps122_buffer,
    stumps_near_school = stumps_near_school,
    stumps_all_districts = trees_with_districts |>
      filter(CounDist %in% comparison_districts, 
             !is.na(stumpdiameter) | tpcondition %in% c("Dead", "Retired"))
  )
  
  return(proposal_stats)
}

# run the analysis
results <- list(stats = analyze_astoria_proposal(trees_with_districts, nyc_districts))
show map code 
# map showing planting opportunities near ps 122q
ggplot() +
  # base layer: district 22 boundary as light gray background
  geom_sf(data = results$stats$astoria_district, 
          fill = "gray95",   # very light fill
          color = "gray50",  # medium gray border
          size = 1) +
  
  # layer 2: existing healthy trees in green
  geom_sf(data = results$stats$district_22_trees_sf |> 
            filter(!(tpcondition %in% c("Dead", "Retired") | !is.na(stumpdiameter))),  # exclude dead/stumps
          aes(color = "Existing Trees"), 
          size = 0.5,    # small dots
          alpha = 0.6) +  # semi-transparent
  
  # layer 3: planting opportunities (dead trees and stumps) in orange
  geom_sf(data = results$stats$district_22_trees_sf |> 
            filter(tpcondition %in% c("Dead", "Retired") | !is.na(stumpdiameter)),  # only dead/stumps
          aes(color = "Planting Sites (Dead/Retired)"), 
          size = 1.5,    # larger dots so they stand out
          alpha = 0.8) +
  
  # layer 4: 500m buffer around ps 122q shown as dashed circle
  geom_sf(data = results$stats$ps122_buffer, 
          fill = NA,              # no fill, just outline
          color = "darkgreen",    # green to match school zone theme
          linetype = "dashed",    # dashed line for "area of focus"
          size = 1) +
  
  # define colors for the two tree categories
  scale_color_manual(
    values = c("Existing Trees" = "forestgreen", 
               "Planting Sites (Dead/Retired)" = "darkorange"),
    name = ""  # no legend title
  ) +
  
  # zoom in to just the buffer area
  coord_sf(xlim = st_bbox(results$stats$ps122_buffer)[c("xmin", "xmax")],
           ylim = st_bbox(results$stats$ps122_buffer)[c("ymin", "ymax")]) +
  
  # clean theme
  theme_minimal() +
  
  # titles and labels
  labs(
    title = "Planting Sites Near P.S. 122Q",
    subtitle = "District 22 (Astoria) - 500m radius shown",
    caption = paste0("Planting sites available: ", nrow(results$stats$stumps_near_school))
  ) +
  
  # customize appearance
  theme(
    legend.position = "bottom",  # legend at bottom so it doesn't cover map
    plot.title = element_text(face = "bold", size = 14)
  )

Planting sites within 500m of P.S. 122Q
show density comparison code 
# prep data for the bar chart - add labels and highlight district 22
density_plot_data <- results$stats$density_comparison |>
  mutate(
    district_label = paste("District", CounDist),  # create readable labels
    is_target = CounDist == 22  # flag district 22 for highlighting
  )

# create horizontal bar chart comparing tree density
ggplot(density_plot_data, aes(x = reorder(district_label, tree_density),  # sort by density (lowest first)
                               y = tree_density,
                               fill = is_target)) +  # color based on whether it's district 22
  
  # create the bars
  geom_col() +
  
  # add text labels showing total tree count for each district
  geom_text(aes(label = format(tree_count, big.mark = ",")),  # add commas to numbers
            hjust = -0.1,  # nudge text to the right of bars
            size = 3) +
  
  # color district 22 orange, others gray
  scale_fill_manual(values = c("TRUE" = "#E69F00", "FALSE" = "gray70"),
                    guide = "none") +  # hide legend
  
  # flip to horizontal bars for easier label reading
  coord_flip() +
  
  # clean minimal styling
  theme_minimal() +
  
  # titles and axis labels
  labs(
    title = "Tree Density Comparison",
    subtitle = "District 22 is one of least dense districts amongst peers",
    x = "",  # no label needed for districts
    y = "Trees per Square Unit",
    caption = "Numbers show total tree count"
  ) +
  
  # adjust text formatting
  theme(
    plot.title = element_text(face = "bold", size = 12),
    axis.text = element_text(size = 10)
  )

Tree density across comparable Queens districts
show stump comparison code 
# bar chart showing stump counts across districts
ggplot(density_plot_data, aes(x = reorder(district_label, stump_count),  # sort bars by stump count
                               y = stump_count,
                               fill = is_target)) +  # highlight district 22 in orange
  
  # create the bars
  geom_col() +
  
  # add text labels showing exact counts
  geom_text(aes(label = stump_count), 
            hjust = -0.1,  # push labels slightly to the right of bars
            size = 3) +
  
  # color scheme: orange for district 22, gray for others
  scale_fill_manual(values = c("TRUE" = "#E69F00", "FALSE" = "gray70"),
                    guide = "none") +  # hide the legend since colors are self-explanatory
  
  # flip coordinates so bars go horizontal (easier to read district names)
  coord_flip() +
  
  # clean theme
  theme_minimal() +
  
  # titles and labels
  labs(
    title = "Available Planting Sites (Stumps)",
    subtitle = "District 22 has significant opportunity for youth engagement",
    x = "",  # no x-axis label needed
    y = "Number of Stumps"
  ) +
  
  # adjust text sizes for readability
  theme(
    plot.title = element_text(face = "bold", size = 12),
    axis.text = element_text(size = 10)
  )

Available planting sites by district
show district map code 
# map showing stumps across all comparison districts
ggplot() +
  # first layer: district boundaries with different colors for each district
  geom_sf(data = results$stats$comparison_district_bounds, 
          aes(fill = factor(CounDist)), 
          alpha = 0.3,        # make them semi-transparent so stumps show through
          color = "gray30") + # darker borders to distinguish districts
  
  # second layer: all the stump locations as red dots
  geom_sf(data = results$stats$stumps_all_districts, 
          color = "red",      # red really pops against the district colors
          size = 0.8,         # small dots since there are many
          alpha = 0.6) +      # slight transparency to see overlap
  
  # use a nice color palette for the district fills
  scale_fill_brewer(palette = "Set2", name = "District") +
  
  # clean minimal theme
  theme_minimal() +
  
  # add titles and explanatory text
  labs(
    title = "Distribution of Planting Sites Across Districts",
    subtitle = "Red points show stump locations (potential planting sites)"
  ) +
  
  # customize theme elements
  theme(
    legend.position = "right",  # legend on the side so it doesn't cover the map
    plot.title = element_text(face = "bold", size = 12)
  )

Spatial distribution of planting sites